Intelligent system for multicriteria optimization of the tehnology plan of two axes turning

Obdelovalni proces je pod stalnim vplivom razvoja in posodabljanja komponent v sistemu stroj-orodje-obdelovanec. Bolj kot so poznana zapletena razmerja med vhodnimi in izhodnimi veličinami, uspešnejši je končni rezultat obdelave s tehničnega in ekonomskega vidika. Obdelava množice medsebojno odvisnih podatkov in omejitev v realnem procesu s klasičnimi metodami ni več primerna, temveč se za optimizacijo poskuša narediti model, ki vključuje določeno stopnjo umetne inteligence. V disertaciji je predstavljen model inteligentnega sistema za samodejno programiranje obdelave struženja na rotacijskih izdelkih. Sistem je zmožen na osnovi 3D-modela izdelka samostojno prepoznati izdelovalnost s postopkom dvoosnega struženja, določiti obdelovalna orodja, zaporedje obdelave, izvesti optimalizacijo in pripraviti NC-program. Nastali NC-program bo glede na izbrane kriterije optimalen glede na čas in učinkovitost obdelave. Razviti inteligentni sistem deluje popolnoma samostojno, brez posega izkušenega strokovnjaka v potek priprave tehnologije. Obdelava se sočasno optimalizira po več kriterijih, po odrezovalnem času, odvzemu materiala, strošku orodja in številu prehodov orodja. Kriteriji se merijo v različnih enotah in se pogosto medsebojno izključujejo, kar pomeni, da izboljšanje obdelave po enem kriteriju pripelje do poslabšanja po drugem. V sistem za samodejno programiranje obdelave je vključena metoda za večkriterijsko optimiranje, analitični hierarhični proces. Metoda temelji razčlenjevanju problema na več ravni, primerjavi kriterijev po parih, izračun vpliva posamezne alternative na končni cilj in preverjanje doslednosti. Rezultati so dodatno preverjeni z metodo za večkriterijsko odločanje ELECTRE II in programom SuperDecisions. Predstavljeni model se lahko uporabi za analizo in optimalizacijo realnih problemov obdelave s struženjem v praksi.Machining processes are consistently affected by the development and modernisation of various components in system machine-tool workpieces. By clarifying the complex relations between input and output values, the final results of processing are more effective from both technical and economic perspectives. Classical optimisation methods are not suitable for processing large amounts of interdependent data in real-world applications, so it is important to develop optimisation models that incorporate a certain level of artificial intelligence. The dissertation presents a novel intelligent system that can program the turning of rotational parts. The system is based on a three-dimensional model of the workpiece. It can independently identify the features involved in turning processes, select cutting tools, develop an operations plan, and perform optimisation. Ultimately, it can produce a numerical control program that is optimised according to selected criteria such as time and productivity. The proposed intelligent system operates entirely independently, without requiring an expert to prepare technological processes. The machining process is simultaneously optimised across multiple criteria: cutting time, material removal rate, price per cutting edge, and number of tool passes. These criteria are expressed in different units and often conflict, meaning that improving processes in terms of one criterion tends to worsen processes in terms of another criterion. The self-programming analytic hierarchy process incorporates methods for multi-criteria analysis: articulating a problem on several levels, conducting pairwise comparisons, ranking alternatives for each decision criterion, and assessing the consistency index. The results were verified using ELECTRE II, a method for multi-criteria analysis, and the SuperDecisions program. The proposed model can be used to analyse and optimise real-world problems involving turning processes